从卫星SAR海洋图像中检测船目标及其航迹

DETECTION OF SHIP AND SHIP WAKE FROM SATELLITE SAR IMAGERY

从卫星SAR海洋图像中检测船目标及其航迹是一项很有意义的工作。由于船目标有强的角反射 ,它的雷达回波信号比较强 ,在SAR图像中的灰度值比较高。先检测船目标 ,再在船的周围寻找航迹的检测方案比较合理。文中先介绍了船目标的检测问题。提出应用非线性拉伸的方法增强船目标 ,运用形态学滤波的方法消除斑点 ,通过取阈值方法检测出船目标后 ,统计船目标的长度以及中心位置等重要信息。检测出船目标后 ,使用曲线扫描的方法来进一步检测航迹特征。文中详细讨论了曲线扫描的扫描中心的确定、扫描半径和扫描步长选取等问题 ,并运用多项式拟合的方法对扫描曲线进行分析 ,给出了航迹特征存在的准则。最后 ,应用实际的SAR图像对算法进行了测试 ,结果表明 ,方案合理。

Detecting ship and ship wake from satellite SAR imagery is significant.Because of strong corner reflection,returned signals from ship are strong and ship in SAR imagery is very bright feature.It would be reasonable to detect ship first rather than ship wake.In the paper we introduce the problem of ship detection.We declare that it is effective to apply non linear stretch to enhance ship objects and to apply intensity morphological operators to eliminate speckles.Thus ships can be identified with few false alarms using simple threshold technique.After that we apapaly curve scan to search ship wake around the ship.We discuss some problems such as determining the center of curve scan,determining the radius of curve scan,determining the step of curve scan and so on.Then we analyze the curve scan data using chebyshev polynomial approximation.The criterion that we use to determine ship wake from curve scan is given in the paper.Finally,we use some satellite SAR imageries to test our scheme.The result shows that te scheme is good.